On 5/14/20 2:55 AM, Damien Le Moal wrote:
On 2020/05/14 9:22, Martin K. Petersen wrote:
Damien,
Any idea why the io_opt limit is not set to the physical block size
when the drive does not report an optimal transfer length ? Would it
be bad to set that value instead of leaving it to 0 ?
The
When building arm32 allyesconfig:
ld.lld: error: undefined symbol: __aeabi_uldivmod
>>> referenced by dm-zoned-target.c
>>> md/dm-zoned-target.o:(dmz_ctr) in archive drivers/built-in.a
dmz_fixup_devices uses DIV_ROUND_UP with variables of type sector_t. As
such, it should be using
On 2020/05/14 11:20, Martin K. Petersen wrote:
>
> Damien,
>
>> Indeed. It is an NVMe M.2 consumer grade SSD. Nothing fancy. If you
>> look at nvme/host/core.c nvme_update_disk_info(), you will see that
>> io_opt is set to the block size... This is probably abusing this
>> limit. So I guess the
Damien,
> Indeed. It is an NVMe M.2 consumer grade SSD. Nothing fancy. If you
> look at nvme/host/core.c nvme_update_disk_info(), you will see that
> io_opt is set to the block size... This is probably abusing this
> limit. So I guess the most elegant fix may be to have nvme stop doing
> that ?
On 2020/05/14 9:22, Martin K. Petersen wrote:
>
> Damien,
>
>> Any idea why the io_opt limit is not set to the physical block size
>> when the drive does not report an optimal transfer length ? Would it
>> be bad to set that value instead of leaving it to 0 ?
>
> The original intent was that
Damien,
> Any idea why the io_opt limit is not set to the physical block size
> when the drive does not report an optimal transfer length ? Would it
> be bad to set that value instead of leaving it to 0 ?
The original intent was that io_opt was a weak heuristic for something
being a RAID
Adding Martin.
On 2020/05/11 22:23, Damien Le Moal wrote:
> On 2020/05/11 20:46, Damien Le Moal wrote:
>> On 2020/05/11 20:25, Hannes Reinecke wrote:
>>> On 5/11/20 12:55 PM, Damien Le Moal wrote:
On 2020/05/11 11:46, Damien Le Moal wrote:
> Mike,
>
> I am still seeing the
On 5/13/20 7:10 PM, Mike Snitzer wrote:
On Wed, May 13 2020 at 1:01pm -0400,
Hannes Reinecke wrote:
On 5/13/20 6:25 PM, Mike Snitzer wrote:
On Wed, May 13 2020 at 7:10am -0400,
Hannes Reinecke wrote:
On 5/13/20 11:41 AM, Damien Le Moal wrote:
On 2020/05/13 16:10, Hannes Reinecke wrote:
On Wed, May 13 2020 at 1:01pm -0400,
Hannes Reinecke wrote:
> On 5/13/20 6:25 PM, Mike Snitzer wrote:
> >On Wed, May 13 2020 at 7:10am -0400,
> >Hannes Reinecke wrote:
> >
> >>On 5/13/20 11:41 AM, Damien Le Moal wrote:
> >>>On 2020/05/13 16:10, Hannes Reinecke wrote:
> Switch to use
On 5/13/20 6:25 PM, Mike Snitzer wrote:
On Wed, May 13 2020 at 7:10am -0400,
Hannes Reinecke wrote:
On 5/13/20 11:41 AM, Damien Le Moal wrote:
On 2020/05/13 16:10, Hannes Reinecke wrote:
Switch to use dynamic debug to avoid having recompile the kernel
just to enable debugging messages.
On Wed, May 13 2020 at 7:10am -0400,
Hannes Reinecke wrote:
> On 5/13/20 11:41 AM, Damien Le Moal wrote:
> >On 2020/05/13 16:10, Hannes Reinecke wrote:
> >>Switch to use dynamic debug to avoid having recompile the kernel
> >>just to enable debugging messages.
> >>
> >>Signed-off-by: Hannes
On 2020/05/13 16:07, Hannes Reinecke wrote:
> Instead of emulating zones on the regular disk as random zones
> this patch adds a new 'cache' zone type.
> This allows us to use the random zones on the zoned disk as
> data zones (if cache zones are present), and improves performance
> as the zones
On 2020/05/13 16:07, Hannes Reinecke wrote:
> Instead of excluding invalid zones we should check if the zone
> is sequential and exclude invalid states. That way we don't need
> to touch the selection when new zone states or flags are added.
>
> Signed-off-by: Hannes Reinecke
> ---
>
On 5/13/20 11:41 AM, Damien Le Moal wrote:
On 2020/05/13 16:10, Hannes Reinecke wrote:
Switch to use dynamic debug to avoid having recompile the kernel
just to enable debugging messages.
Signed-off-by: Hannes Reinecke
---
include/linux/device-mapper.h | 7 +--
1 file changed, 1
On 2020/05/13 16:07, Hannes Reinecke wrote:
> Hi all,
>
> here are two patches to improve performance when a regular device
> is specified for dm-zoned.
> Damien noted that we have a performance drop after the zones on the
> regular device are used up, as then the random zones on the zoned
>
On 2020/05/13 16:10, Hannes Reinecke wrote:
> Switch to use dynamic debug to avoid having recompile the kernel
> just to enable debugging messages.
>
> Signed-off-by: Hannes Reinecke
> ---
> include/linux/device-mapper.h | 7 +--
> 1 file changed, 1 insertion(+), 6 deletions(-)
>
> diff
Hello Hannes Reinecke,
The patch 70978208ec91: "dm zoned: metadata version 2" from May 11,
2020, leads to the following static checker warning:
drivers/md/dm-zoned-target.c:1047 dmz_iterate_devices()
warn: was expecting a 64 bit value instead of '~(zone_nr_sectors - 1)'
Switch to use dynamic debug to avoid having recompile the kernel
just to enable debugging messages.
Signed-off-by: Hannes Reinecke
---
include/linux/device-mapper.h | 7 +--
1 file changed, 1 insertion(+), 6 deletions(-)
diff --git a/include/linux/device-mapper.h
Instead of emulating zones on the regular disk as random zones
this patch adds a new 'cache' zone type.
This allows us to use the random zones on the zoned disk as
data zones (if cache zones are present), and improves performance
as the zones on the (slower) zoned disk are then never used
for
19 matches
Mail list logo